An Abstraction Layer to Facilitate Technical Interoperability Between Medical Records and Knowledge Modules

Stud Health Technol Inform. 2017:243:185-189.

Abstract

Integrating clinical decision support (CDS) functions into an existing hospital information system (HIS) is often a tedious task. This problem area is so pervasive that the Arden Syntax, a widely used standard for CDS functions, assigned a specific designation, the so-called "curly braces problem". It derives from a pair of curly braces used to encapsulate any parameters required for the interactions with a HIS. The traditional approach is to leave the problem area of technical interoperability entirely to the specific institution, possibly entailing a considerable amount of initial programming work. This study describes a reusable and expandable solution to this problem in the form of an abstraction layer. Our study comprised an analytical phase in which we investigated the data source access capabilities of five Arden Syntax environments. Building on the results, we implemented a working prototype that is capable of querying heterogeneous data sources, which facilitates a straightforward connection of new data sources with existing and future communication protocols and standards. From our point of view, the technical aspects of the "curly braces problem" with respect to data source access have changed over the years, insofar as technical progress lead to defacto standards for data storage, data querying and inter-system communication. An agreement on such a convention, together with the supply of commonly used data source adapters could promote the further dissemination of the Arden Syntax as a standard for representing and sharing medical knowledge.

Keywords: Arden Syntax; Clinical decision support; curly braces problem; interoperability.

MeSH terms

  • Artificial Intelligence
  • Decision Support Systems, Clinical*
  • Electronic Health Records*
  • Hospital Information Systems*
  • Humans
  • Information Storage and Retrieval*
  • Knowledge Bases
  • Medical Records
  • Programming Languages